New neural network types estimating the accuracy of response for ecological modelling

Modeling water and carbon fluxes above summer maize field in North China Plain with back-propagation neural networks

“…Artificial neural networks (ANNs) are attractive and promising strategies for this work because of their capacity in prediction, control and optimization of input-output responses without a predefined mathematical model (Kosko, 1992;Demuth and Beale, 1994;Schulz and Härtling, 2003) in many research fields. A few applications of this method had been reported in modeling of ecological data since the beginning of the 90's (Thai and Shewfelt, 1991;Chao and Anderson, 1994;Murase et al, 1994;Cook and Wolfe, 1991;Elizondo et al, 1994;Batchelor et al, 1997;Lek et al, 1996;Lek and Guegan, 1999;Hecht-Nielsen, 1987;Huntingford and Cox, 1997;Francl and Panigrahi, 1997;Werner and Obach, 2001;Moisen and Frescino, 2002). Most of these works showed that ANNs performed better than classical modeling methods.…”

Proactive management of estuarine algal blooms using an automated monitoring buoy coupled with an artificial neural network

“…ANNs learn the input/output relationship through training which can be unsupervised or more commonly supervised. Thus ANNs have the potential to create models even when the underlying domain knowledge of ecosystem processes, and their interaction, is limited (Barciela et al, 1999;Werner and Obach, 2001;Wilson and Recknagel, 2001). Once the ANN is properly trained, it has learned to model the unknown function that relates the input variables to the output variables, and can be used to make predictions when the output is not known.…”

Hydrological and Hydraulic Aspects of Restoration Uncertainty for Ecological Purposes

“…Alternatively, physically-parameterised models may be replaced by stochastic or hybrid approaches such as cellular autometa (Chen et al, 2002) or neural network models (Werner and Oback, 2001;Reyjol et al, 2001;Gevrey et al, 2003) as witnessed in hydrology and fl oodplain inundation studies. As yet, however, these approaches have focused on simulating ecological characteristics: appropriate scales for applications, for the development of rules and training data sets remain to be explored.…”